Static and Dynamic Views of European Integration

There is a theoretical and empirical need to distinguish between static support for the EU as it now is and dynamic support for further integration. Although most Europeans endorse the EU as a good thing today, the European Election Study finds no popular majority for an ever closer union, the commitment of EU institutions. Less than one‐third endorses further integration and less than one‐third thinks integration has gone too far. The largest group favours keeping the EU as it is. Their outlook reflects ambivalence; they see the EU as having both strengths and weaknesses. It does not reflect lack of EU knowledge or of socio‐economic resources, as is the case with ‘don't knows’. While eurozone institutions are committed to further integration, most EU citizens are not. Likewise, there is no majority supporting eurosceptic demands for returning powers to national governments.


Introduction
The concept of European integration is ambiguous: it has both static and dynamic meanings. Studies of public opinion normally ask people to evaluate the European Union as it is today (see Hobolt, 2012). This leaves open the opinions that respondents have about change in any direction. A dynamic definition, stated explicitly in the Treaty of Rome, is that European integration is a process of movement toward an ever closer Union. However, this is not the only form of change that is possible. Soft eurosceptics make the case for less integration through the repatriation of EU powers to national governments and hard eurosceptics argue for change in the form of their country exiting from the Union (see Leconte, 2010). The dynamics of the eurozone are often conceptualized in terms of a dichotomous choice between greater integration or the disintegration of the Eurozone, or even the European Union (see e.g. Hodson, 2013).
Logically, current and dynamic views of Europeans can be combined in four different ways. People who endorse the EU in its current form can also favour further measures of integration; this is the position of the leaders of EU institutions. Alternatively, people who dislike the current state of the EU can seek less integration, the stated position of the British government under David Cameron and the German Alternative for Deutschland, or they can want their country to withdraw from the EU, the stated position of the French National Front (www.EU&I.eu). For people whose current and future preferences are in harmony, the theoretical distinction between a time-specific and a dynamic view is of no consequence: knowing a person's current evaluation tells you their view of further integration.
People who positively evaluate the EU as it is today but do not endorse further integration combine contrasting views: they are conservative in the literal sense, preferring stability to change. Herbert Simon's (1978) theory of satisficing justifies describing this group as satisfied with the status quo. The position may be justified by being positive about existing institutions or sceptical about the putative benefits of any change. A preference for no change is consistent with Kahnemann and Tversky's (1979) theory of individuals being risk averse. Avoiding the risks of change is an argument often used against Greek withdrawal from the eurozone and British exit from the EU. Dissatisfaction with the EU as it is today and wanting more integration is the EU's formal response to the eurozone crisis (Tuori and Tuori, 2014). Democratic critics express their dissatisfaction with the status quo by supporting the reform of EU institutions of representation in a politically more integrated Europe (cf. Bellamy, 2012;Laffan, 2014).
Since the Treaty of Rome came into effect, institutional dynamics have promoted further integration. Intergovernmental treaties conferring additional powers on European institutions have been negotiated between national governments (Moravcsik, 1998;Pollack, 2012), without what Haas (1968, p. 17) and Moravcsik and Schimmelfennig (2009, p. 84) have described as 'unnecessary' public opinion surveys or public debates. The cumulative effect of small-scale increases in powers without political negotiations (Dehousse, 2011;Sandholtz and Stone Sweet, 2012) has been described as relying on the public's passive 'permissive consensus' to allow 'integration by stealth' (Haas, 1968, p. 17;Lindberg and Scheingold, 1970;Majone, 2005). However, the progressive increase in the European Union's visible influence on the lives of Europe's citizens has had its own spillover effect: 'The times when elites could pursue European integration with no regard to public opinion are long gone' (Hobolt, 2012, p. 716).
This article is innovative in demonstrating both the theoretical and empirical importance of taking into account Europeans who want to maintain the status quo when analysing popular support for an ever closer Union. The next section shows that the median and largest group favours keeping the EU as it is, rather than endorsing more integration, endorsing less integration or having no opinion. Five hypotheses are offered to account for these differences. Multivariate statistical tests find that evaluations of government performance are most important for discriminating between those wanting more or less integration, while ambivalence about performance is most important for those in favour of the status quo. Lack of resources best explains the 'don't knows'. The conclusion calls attention to implications for the EU's institutionalized commitment to increasing political integration without the unqualified commitment of two-thirds of its citizens.

I. Attitudes Toward European Integration
A substantial body of research has conceptualized and measured popular evaluation of the European Union in a wide variety of ways (Loveless and Rohrschneider, 2013). In a substantial review of the literature, Boomgaarden et al. (2012: 245ff) identified five different sets of attitudes toward the EU. Statistical analysis of a multi-item Dutch survey identified one of these as a dynamic factor labelled 'strengthening'. The five items are conceptually distinct, concerning everything from categoric approval of a United States of Europe to enlargement of the EU; the notional integration factor is the statistically weakest of the five (Boomgaarden et al., 2012 Table 1; cf. Anderson and Hecht, 2012: 17ff).
The standard Eurobarometer measure of support for integration is static: Generally speaking, do you think that your country's membership of the European Union is a good thing, a bad thing, or neither good nor bad? Although four alternatives are offered, a clear majority usually evaluates the EU today as a good thing (Leconte, 2010, p. 164;Henjak et al., 2012). Since this static question is regularly asked in the Eurobarometer, it has been the object of dozens of political science analyses that dichotomize replies into those thinking the EU is a good thing and a composite category of those saying neither good nor bad, bad, or don't know (Hobolt, 2012, p. 517ff). After a sophisticated time-series econometric analysis, Franklin and Wlezien (2013, p. 6) conclude that while the short-term shock of events causes current evaluations to fluctuate, there is a long-term stable equilibrium of about 60 per cent thinking the EU is a good thing. Since EU integration has increased every substantially over the decades spanned by the Eurobarometer (Dinan, 2010), a steady level of support is often interpreted in Stimson's (1991) terms as showing a 'zone of acquiescence', an after-the-fact endorsement of steps that Europe's policy-makers have taken toward an ever closer Union.  Romania  54  19  12  15  Spain  51  31  14  4  Poland  46  32  10  13  Slovenia  42  30  26  2  Greece  41  28  28  3  Italy  40  27  25  8  The Netherlands  39  40  19  1  Cyprus  36  28  30  7  Germany  36  36  27  1  Denmark  35  43  22  1  Lithuania  33  28  16  23  Sweden  33  43  23  2  Czech Republic  32  35  24  10  France  32  32  33  4  Luxembourg  32  35  30  2  Malta  32  24  15  30  Bulgaria  31  31  14  24  Ireland  31  48  20  2  Slovakia  31  35  15  18  Belgium  29  39  25  7  Hungary  29  31  27  13  Portugal  27  34  22  17  Austria  25  36  37  2  Estonia  25  32  32  11  Finland  19  48  32  1  United Kingdom  16  41  41  3  Latvia  15  27  44  14  Mean  33  34 25 9 Standard Deviation 9 7 9 8 Source: European Election Study 2009; Q80 recoded as follows: 0-3 anti-integration, 4-6 satisfied as is, 7-10 pro-integration; N=27,069.
However, as the EU's impact on national policies has increased since the Maastricht Treaty, there has been a shift from passive acquiescence in further integration to a 'constraining dissensus' reflected by the rise of eurosceptic parties (Down and Wilson, 2008;Hooghe and Marks, 2009). The blanket term 'eurosceptic' lumps together parties that differ in whether they are parties that are against further integration, parties that want to reduce EU powers, and parties that favour withdrawing from the EU altogether (Leconte, 2010, chapters 9-10;Lubbers and Scheepers, 2010). Some studies conclude that the prime purpose of so-called eurosceptic parties is to protest against their national governors rather than against what is done in Brussels (cf. Szczerbiak and Taggart, 2008;Mudde, 2013;Ford and Goodwin, 2014). Focus group discussions find many examples of individuals being ambivalent, associating the European Union with both good things and bad things (Duchesne et al., 2013). From such studies, van Ingelgom (2014, p. 176) concludes, 'An interpretation based on the binary of permissive consensus and euroscepticism is empirically incomplete if not erroneous'.

A Fourfold Division of Dynamic Evaluations
To assess popular attitudes, a question is needed that recognizes that integration is a dynamic process because, as Wessels (1995;145) emphasizes, 'the willingness to use the political and economic resources of one's own country to deepen integration demands a more active degree of support'. While the Eurobarometer (2013) routinely collects data evaluating the European Union as it is, it does not couple this with a question about support for the principle of further integration. Consistent with the interests of DGs wanting to expand their particular activities, it may ask about support for specific policies that would incidentally have a spillover effect on integration.
The EES (European Election Study) explicitly measures dynamic attitudes toward integration (for details, see www.piredeu.eu). In its 2009 post-election survey interviewing 27,069 people, it asked: Some say European unification should be pushed further. Others say it has already gone too far. What is your opinion? Respondents are offered an 11-point scale in which Integration has already gone too far is the 0 point and point 10 is Integration should be pushed further. The description of point 10 avoids the emotive symbolism of a United States of Europe. Likewise, referring to integration as having 'gone too far' avoids the complementary extreme of proposing national withdrawal or the breakup of the European Union. Those who choose the intermediate location, point 5, cannot be described as either pro or anti-integration, because they have rejected both of these alternatives. Since the 2009 EES study was the first to ask this appropriate measure of the dynamics of integration, the following analysis provides a base line for evaluating change since then (see Figure 1).
By contrast, the use of the 1-10 scale used in the 1999 and 2004 European Election Studies did not offer a clear opportunity to show satisfaction with the status quo. A reply of '5' is mathematically part of the negative half of the scale, but psychologically many respondents see it as an intermediate category between favouring more or less integration. In analysing these surveys, Toka et al. (2012) simply treated the scale as a measure of the degree to which respondents endorse further integration. McLaren (2006, p. 27ff) used a dynamic Eurobarometer question from 2000 that asked whether people wanted the EU to play a more important, the same, or a less important role in their daily life. However, the dependent variable she created reduced its dynamic utility by combining this question with a static measure of whether EU membership is a good thing or a bad thing.
Dynamic attitudes toward European integration are tri-modal, not bi-modal ( Figure 1). The largest group, 21 per cent, explicitly reject commitment to either more or less integration by placing themselves at 5, the mid-point of the scale. Doing so indicates satisfaction with the status quo by comparison with change (cf. Simon, 1978). Those favouring the status quo are almost twice as numerous as each group at opposite ends of the scale. Since those placing themselves at points 4 and 6 are much closer to the scale's mid-point than either end point, consistent with spatial theories of individual preferences (Enelow and Hinich, 1984), we group them among those satisfied with the status quo. Since those at points 3 and 7 clearly position themselves with the anti-and pro-integration groups, they are assigned to these respective categories. By refusing to place themselves anywhere on the scale, the 'don't knows' constitute a fourth category. On this basis, 34 per cent of EES respondents are satisfied with the EU as it is; 33 per cent favour further integration; 24 per cent think integration has gone too far; and 9 per cent are 'don't knows'.
The fourfold division of opinion is found in every Member State (see Table 1). In the great majority of countries the median citizen is satisfied with the EU as it is, and in 15 countries the satisfied are the largest group. Only two Member States -Romania and Spainshow a majority in favour of further integration, while only in Latvia and the United Kingdom do as many as two in five think European integration has gone too far. In 12 new Member States where citizens had had less than six years of exposure to EU membership, an average of 15 percent were 'don't knows', compared to 6 per cent in the six founder states.
Comparing static and dynamic views about integration demonstrates that the two questions do not measure the same opinion of individuals, because a majority change their position when asked to evaluate future integration (see Table 2). Consistent with the Eurobarometer long-term average, the EES survey found that 63 per cent viewed the EU as a good thing. However, among those seeing the EU as a good thing, 57 per cent do not want integration to go further. Among the much smaller group seeing European integration as bad, more than three-fifths think it has gone too far, while one-quarter are prepared to accept the status quo or tolerate further integration. 1 Among those who say the EU is neither good nor bad, more than four-fifths withhold endorsement of further integration. In sum, only 34 per cent of 2009 EES respondents conform to the stereotype assumption that those seeing the EU as a good thing want more of the same and those seeing it as bad favour less integration. Thus, to avoid misunderstanding, a static measure cannot be used as a proxy for dynamic preferences; it is necessary to use a dynamic measure.

II. Explaining Dynamic Preferences
To explain dynamic preferences toward more or less European integration, we need a model that takes into account four different alternatives. We also need to include a multiplicity of influences in our model, since what affects being for or against integration may be different from influences on endorsement of the status quo or having no opinion. To account for the fourfold empirical distribution of European opinion today, we set out hypotheses that differ in which attitude they seek to explain, as well as in the theoretical influences that they postulate. Since individuals have formed their EU evaluations in 27 countries that differ nationally (see Table 1), national context as well as individual characteristics is likely to be significant.
A positive evaluation of the performance of government provides a utilitarian justification for being committed to further integration (Scharpf, 1999). However, this leaves open whether it is the performance of national or of EU government that people evaluate. Reif and Schmitt (1980) postulated that citizens tend to project onto the EU judgements of their national government, whether favourable or unfavourable. However, the EU's accumulation of powers that operate concurrently with national government has led Hooghe and Marks (2009, p. 14) to conclude, 'The European Union is no longer insulated from domestic politics; domestic politics is no longer insulated from Europe'. For example, immigration, which is part of the EU's promotion of the free movement of people, can affect people in countries that attract immigrants. Especially in eurozone countries, economic contraction and the dissatisfaction with the national government that it stimulates may be blamed on the EU as well as or instead of national government. The multi-level system of governance in Europe today makes it increasingly difficult for ordinary citizens to 1 The fourth logical category, those who think the EU is a bad thing that would be improved by further integration, comprise only 1 per cent of EES respondents. distinguish whether such major political concerns as the economy or immigration reflect the performance of national or EU institutions, or both (Duchesne et al., 2013). Empirical research on voting at EP elections has confirmed that both national and European influences are important (see e.g. Hix and Marsh, 2011). Our performance hypothesis leaves open to empirical determination the extent to which national or EU government is the source of influence.
H1 PERFORMANCE: If individuals evaluate the performance of government positively, they are more likely to favour further rather than less integration.
Engagement is frequently theorized to encourage individuals to favour integration. Engagement can involve emotional affect, behaviour or knowledge. Identifying with Europe is the major form of emotional engagement, while those who identify only with their nation should be more likely to see integration as having gone too far (cf. Risse, 2010;Kaina and Karolewski, 2013). Behavioural activities, such as voting in a European Parliament election, watching EP election news and the correlate, knowledge of the EU, are likewise deemed to re-enforce support for further integration. At the contextual level, Inglehart (2008) has theorized that the longer a country has been an EU Member State, the greater the proportion of its citizens who have been socialized for much or all their lives to view the EU as part of their political system and thus support further integration. However, eurosceptic campaigns assume that the more one knows about the EU, the less support there will be for further European integration (Leconte, 2010: chapter 6). Since the overwhelming majority of Europeans are first socialized into national politics, interest in national politics and having a political ideology, whether of the left or the right, may also affect attitudes toward Europe.
H2 ENGAGEMENT: If individuals are politically engaged, they are more likely to favour further rather than less integration.
From its founding as the European Economic Community, the European Union has promoted economic integration. Economic integration distributes both benefits and costs. Fligstein (2008) has theorized that the Europeanization of economic life chances has created a division within national societies between winners and losers (see also Kriesi et al., 2012). Winners are defined as the more educated, middle-class and higher income groups with the socio-economic resources to benefit from further integration, while losers are those who lack the resources to adapt to increased economic integration and are therefore more likely to think it has gone too far.
H3 WINNERS AND LOSERS: If individuals have more socio-economic resources to benefit from integration, they are more likely to favour further rather than less integration.
The attitude of people who reply 'don't know' when asked about further European integration is appropriately explained by theories of public opinion that emphasize a lack of socio-economic resources as causing individuals to have no opinion about major political issues (Dalton and Klingemann, 2007). Inglehart (2008) notes that this is especially relevant to opinions about the European Union, since its institutions are most distant from the everyday lives of people; therefore, understanding the EU requires more education and interest in politics. Contrary to Fligstein (2008), who views limited resources as stimulating anti-European attitudes, people low in resources can simply be political drop-outs.
H4 LACK OF RESOURCES: If individuals are low in socio-economic resources, they are more likely to have no opinion about European integration.
Individuals who are ambivalent do not adopt a holistic view of European integration as either good or bad. Ambivalent Europeans discriminate, evaluating some features of the EU positively and others negatively. Having two views of Europe is consistent with European integration having two meanings, one referring to the EU as it is today and the other referring to further integration. For example, a Briton can evaluate the EU as a good thing today, while not wanting further integration through the UK joining the eurozone.
The intermediate position in the static EU integration questionthat is, regarding it as 'neither good nor bad'is often treated as a mark of indifference or ignorance (cf. Nilson, 2002;Mutz, 2002). However, holding conflicting attitudes may reflect an informed and balanced view (Steenbergen and de Vries, 2012;Stoeckel, 2013). For ambivalent individuals, the constraining dissensus about European integration identified at the national level exists within their own mind. It is met by the EES question offering ambivalent respondents an alternative to appearing to have no opinion. They can endorse a position that strikes a balance between conflicting influences pulling them in opposite directions. Refusing to give an unqualified endorsement to change in any direction can reflect satisfaction with the status quo.
H5 AMBIVALENCE: If individuals hold conflicting political attitudes, they are more likely to favour the EU status quo rather than endorse change.

III. Testing Hypotheses About Integration
The theories reviewed above differentiate European citizens into two groups that appear to be mirror images of each otherthose who favour further integration and those who think it has gone too farand two whose relation to these mirror images is problematic. Therefore, we use multinomial logit in STATA (Rabe-Hesketh and Skrondal, 2008) because it makes no assumption about whether there is a linear relationship among the four groups constituting our dependent variable, attitudes about integration. Instead, this form of logit separately identifies significant influences for each group when compared to the reference group, those favouring further integration. Given that the EES survey covered 27 EU Member States, the analysis is multi-level, with contextual indicators clustered. Although context indicators are statistically analysed at a separate level, in Table 3 we present them in relation to the hypothesis for which they are relevant. To determine statistical significance, <.001 is our cut-off point for individual variables and <.10 for contextual variables. Odds ratios indicate the strength of significant influences. (Details of the coding of all variables are given in the Appendix). An odds ratio of more than 1.00 indicates that an independent variable makes people more strongly committed to their position than those who are pro-integration. An odds ratio of less than 1.00 indicates that an independent variable makes people less committed than those who are pro-integration. 2

Divisions Between Pro and Anti-Integration Groups
As hypothesis 1 predicts, how people evaluate government performance is of major importance in differentiating those favouring more and less integration ( Table 3). Evaluation of the EU today is measured by a five-point scale combining replies to two EES questions: trust in EU institutions (47 per cent have trust) and whether EU decisions are usually in the best interests of your country (44 per cent are confident of this). 3 As expected, the more positive people are about the EU, the more strongly they endorse further 2 To test whether different results would be obtained by assigning 4s and 6s to the anti and pro-integration groups respectively, we re-ran the logit in Table 3 with this adjustment. The statistical results were similar, and so was the identification of independent variables as significant or not significant. The only difference is consistent with our initial decision to group replies by spatial distance. By assigning the 4s and 6s to extreme groups this made the two categories more moderate, thus reducing slightly the size of odds ratios representing the difference between the pro and anti groups by a few decimal points. 3 . The tau beta correlation between these two variables is 0.36, significant at the .000 level. integration. The perceived impact of the EU on national life is measured by combining replies to four questions about its effect on the national economy, interest rates, immigration and health care. The mean response is that it has some impact. People who see the impact of the EU as stronger are more likely to favour further integration.
The dichotomy between national politics and EU politics proposed by Reif and Schmitt (1980) has been undermined as competences for immigration and the management of the economy have been integrated in a multi-level system of governance. The EU's single European market allows the free movement of people between Member States and enlargement has made immigration more attractive by increasing economic disparities between Member States. A total of 57 percent of EES respondents think that immigration has gone too far and the effect of this view substantially increases the belief that integration has gone too far (odds ratio 1.41). The change in gross domestic product between 2008 and 2009 involved economic contraction rather than growth, thus challenging the putative benefit of the EU's economic policies. The more the economy has contracted, the less likely people are to favour further integration, the policy the EU has adopted to deal with the economic crisis. The relevance of the EU to economic performance is underscored by the failure of dissatisfaction with the national economy to have a statistically significant effect.
Dissatisfaction with the performance of national government pushes in two opposing directions, depending on the cause. In EU Member States that are rated more negatively on the Transparency International Corruption Perception Index (www.transparency.org), citizens are significantly more likely to favour further integration. This appears to reflect an expectation or hope that greater EU influence will reduce national malgoverno. The 52 per cent generally dissatisfied with their national government tend to project their feelings onto EU governance, thus making them more likely to be against further integration.
Engagement with Europe has a significant effect on attitudes toward further integration but not always in the direction predicted in hypothesis 2. Consistent with a large body of social science theory, individuals who identify themselves as Europeans are significantly more likely to favour further integration. However, the evidence rejects the assumption of Brussels policy-makers that greater knowledge of Europe will lead to greater support for integration. The opposite is the case: knowing more about the EU significantly increases the view that integration has gone too far (odds ratio 1.17). This finding is a reminder that greater knowledge is likely to bring with it more knowledge of the EU's deficiencies as well as its strengths. Likewise, familiarity with the EU as a result of living in one of the six founder states makes an individual more likely to think integration has gone too far, while those with least exposure because they live in new Member States are significantly more likely to be pro-integration.
Two measures of engagement with EU affairsvoting in a European Parliament election and watching news of that election on televisiondo not differentiate supporters and opponents of integration. Two more measures of political engagement relevant to national as well as EU levelsinterest in politics and seeing oneself as on the left or the rightlikewise have no significant influence. The failure of the engagement hypothesis to gain substantial support indicates that those who reject an ever closer Union are just as engaged with European affairs as those who endorse integration.
Notwithstanding the impact of the European Union on ordinary Europeans, it has not created the trans-national division between winners and losers anticipated in hypothesis 3.
Those who hold opposing views for and against integration do not differ significantly from each other in their social class, education, age or gender (Table 3). The only significant influence is an individual's self-assessed standard of living; the lower it is, the more likely a person is to think integration has gone too far. The weak influence of socioeconomic differences is consistent with the Kinder and Kiewiet (1979) theory that egocentric judgements emphasized by winners-and-losers theories are less important than judgements of the collective performance of government.

'Don't Knows' in a Class of Their Own
Consistent with hypothesis 4, those with no opinion about European integration are low in socio-economic resources ( Table 2). 'Don't knows' are the only group whose attitude is significantly influenced by education and age; they are less educated and older than other Europeans. Having a lower standard of living has a bigger effect on the don't knows than on other groups and they are also more likely to be women.
'Don't knows' are also less engaged in politics. They have great difficulty in placing themselves on the conventional left/right scale (odds ratio: 2.37). This is not because they have strong political views that are orthogonal to that scale, but because they are much less interested in politics generally. As for the EU, 'don't knows' are less knowledgeable and opinionated, and since they are much more likely to live in new Member States, they have not been fully socialized into identifying with Europe. 'Don't knows' are significantly less likely to be positive about Europe. However, this is not a judgement on EU performance; it is due to the 'don't knows' being less likely to have any view. For each of the questions making up the evaluation of the EU indicator, the 'don't know' group is four times more likely than others to lack a positive or a negative opinion. While this profile of the 'don't knows' is consistent with the negative stereotype that EU elites often use to reject increased popular participation by uninformed citizens, it applies to only 9 per cent of EES respondents (Figure 1).

Ambivalent and Satisfied
The ambivalence hypothesis postulates that those who are satisfied with the status quo are subject to a combination of influences pushing them in opposite directions, some favouring more and some favouring less integration. Multinomial logit makes it possible to test this hypothesis by identifying influences that significantly affect those wanting no change by comparison with other categories of EES respondents. If those we group as satisfied were inclined to show permissive consent to further integration, a lot of influences would significantly differentiate them from the anti group, but not from those favouring further integration. However, if the satisfied are inclined toward less integration as a second choice, then influences would significantly differentiate them from those wanting to go further, but not from those thinking integration has already gone too far. If the choice of an intermediate answer is really a form of concealing no opinion, then the multinomial logit should show many of the same influences affecting both those we labelled satisfied and the 'don't knows'. A balanced form of ambivalence is that instead of veering between the two extremes, influences usually place respondents in this group between those with pronounced pro or anti-integration views.
The largest group of EES respondents, those satisfied with things as they are, have a balanced view of the integration process. In most cases in which performance or engagement variables exert a significant influence, the odds ratios place them between the pro and anti-integration blocs (Table 3). Among the significant performance influences, being positive about the EU has an odds ratio of 0.45 for those who are anti-integration compared to 0.74 for the satisfied and the implied 1.00 for those favouring more integration. Similarly, for those thinking that immigration has gone too far, the odds ratio differentiating the anti-integration group from the further integration group is 1.41 while the odds ratio for the satisfied is 1.11. The effect of economic contraction significantly differentiates proponents of the status quo from advocates of further change, but it is not as strong as the effect on the anti-integration group.
To describe the satisfied as being in two minds about European integration is correct, but need not imply that individuals feel strong internal conflicts. In terms of engagement with politics, the satisfied and the anti-integrationists are less likely to identify with Europe, but their rationales differ. For the anti-integration group, not having a European identity appears to be part of a syndrome of nationalist attitudes that reject many features of the EU as it is today. For the satisfied, not thinking of oneself as a European is a sign of less strong engagement with politics. The satisfied are significantly less likely to be interested in politics than either the anti or the pro-integration groups. Similarly, while both pro and anti-groups are alike in placing themselves on a left/right scale, those who are satisfied are significantly less likely to think that this scale reflects their political outlook. Being less engaged with politics is also a mark of the 'don't knows', but unlike the latter group, those who are satisfied are not uninformed, non-participating and lacking in education. The satisfied are just as likely to vote in a European Parliament election and watch EP election news as those who are firmly pro or anti-integration. Moreover, the satisfied are just as likely to be knowledgeable about the European Union as those who are pro-integration and significantly more likely than pro-integrationists to live in an EU founder state.

Conclusions
The theoretical importance of the often overlooked distinction between static and dynamic support for the European Union is empirically confirmed. Our fourfold classification shows that the debate about the future of the EU cannot be reduced to a dichotomous division between those wanting more and less integration. The median group of European citizens are not committed to change; they prefer the status quo.
Of the three sets of influences on attitudes to integration, the performance of European political institutions has the most effect. Having a positive view of the EU's performance and awareness of its impact are of first-order importance for dynamic preferences. The effect of the EU is also registered through individuals' negative evaluations of the EU's impact on immigration and the economy. As for engagement, although diffuse identification with Europe does boost support for further integration, the opposite is not the case. Non-identifiers are as likely to endorse the status quo as to think integration has gone too far. Specific measures of engagement such as voting at European elections or watching EP election news have no significant influence on attitudes. Contrary to the assumption of well-funded EU communications programmes that more information about Europe will create more support for integration, knowledge of the EU and living in an EU founder state are significantly more likely to encourage the view that integration has gone too far. Although European integration has undoubtedly created both winners and losers within many societies, being more educated, a major asset for success in a cosmopolitan Europe, has no significant effect on views. Nor does subjective social class, a measure of perceived life chances. There is partial support for hypothesis 3 in that lower income people are less likely to favour further integration. There remains the possibility of an indirect effect too. At national elections, individuals who see themselves as losers may vote for protest parties that, among other things, are anti-EU.
Since the 2009 EES, there has been time for independent variables to change, and with those, attitudes toward further integration. However, variables having a significant influence on attitudes toward integration differ in their potential for change. Major components of political engagement, such as the length of time a country has belonged to the EU, are givens and the percentage of the population identifying themselves as Europeans requires a decade or generation to alter substantially (cf. Inglehart, 2008;Down and Wilson, 2013). Most socio-economic characteristics are not subject to short-term change and in any case are not significant influences. Performance variables are most immediately subject to alteration in response to political and economic change. In the single European market, economic growth targets have failed to be met. The eurozone crisis has had negative effects on a substantial portion of Europeans and the eurozone media have been ready to link national economic issues with the EU.
The post-2014 European Parliament election survey (www.eeshomepage.net) confirms that there has been a change in the expected direction in attitudes toward further integration. However, the extent is marginal: 8 per cent fewer Europeans now endorse further integration and 8 per cent more Europeans now think that integration has gone too far. The largest group remains the same: the 34 per cent who on balance prefer the EU not to change in either direction. The satisfied group has remained the largest because the minority shifting from ambivalence to anti-integration has been replaced by those who previously favoured further integration becoming ambivalent. See Figure 2. 4 The limited shift in public attitudes toward European integration is consistent with our theoretical approach, which emphasizes the importance of a multiplicity of influences on attitudes. The standard question about whether the EU is today a good or a bad thing showed a parallel marginal change; the majority endorsing the EU as a good thing fell by 9 percentage points from 63 per cent in 2009 to 54 per cent in 2014. Substantial changes in the EES questionnaire between 2009 and 2014 make it impossible to evaluate the extent of change in major performance variables. One indicator of being positive about the EU -thinking it makes decisions in the interest of one's countrywas dropped and the coding of the other, about trust in the EU, was altered.
The response of European institutions to the eurozone crisis has been a process described in French as la fuite en avant, 'running away forwards'that is, the adoption of political and economic measures that increase integration. The European Council communique of 27 June 2014 that approved Claude Juncker as the new President of the European Commission reaffirmed the commitment to an ever closer Union. In an associated press conference, the German Chancellor Angela Merkel acknowledged that while Member States might progress at different speeds, this will not stop movement toward further integration (Shipman and Pancevski, 2014).
In public policy terms, the above evidence shows that while EU institutions are moving toward an ever closer Union, European citizens are not. The largest group are in favour of the EU as it is, and only a quarter to a third endorse further integration. Given this, EU officials do not want to consult with public opinion through such means as referendums on treaties. The former President of the European Commission, Jose Manuel Barroso, has justified this position by saying that direct consultation with EU citizens was undesirable because it would require 'simplifying important and complex subjects' for decision by people uninformed about EU affairs (quoted in Hobolt, 2009, p. 23). The 2014 European Parliament election showed an increase from one-sixth to almost one-third in the number of MEPs elected by anti-EU parties. Nonetheless, two-thirds of MEPs have been elected on pro-integration programme (see Borz and Rose, 2010;Rose and Borz, 2013;www.EUandI.eu). However, the principle of subsidiarity gives citizens the opportunity at national elections to hold their government accountable for what they accept at the EU level (Rose, 2014).
Unlike a public opinion survey, there is no intermediate choice in a national referendum ballot on EU measures. The ambivalence of so large a portion of Europeans shows that the swing bloc of voters is not firmly committed against further integration, just as it is not committed in favour. However, to get this group to endorse any change would require clear and convincing arguments showing that the EU's performance would be improved by further integrationarguments that eurosceptics can challenge. At the end of a referendum campaign, the default position of satisfied conservatives is to endorse the status quo. If the ballot proposal is to increase current powers, the likelihood is a vote against an ever closer Union. However, if a referendum ballot proposes withdrawal from the EU, the default position is to vote to remain a Member State.